Landslide susceptibility mapping based on K-Means and Self-Organizing Map clustering with Geographic Information System in Tasikmalaya, West Java

Landslides are one of the most frequent natural disasters in Indonesia, primarily caused by complex topographic conditions, high rainfall intensity, and extensive land use changes. This study aimed to map landslide-susceptibility areas in Tasikmalaya Regency, West Java, using the K-Means Clustering and Self-Organizing Map (SOM) methods, visualized through a Geographic Information Systems (GIS). The data utilized include Landsat 8 satellite imagery for calculating the Normalized Difference Vegetation Index (NDVI) and Normalized Difference Built-up Index (NDBI) indices, elevation and slope data derived from Digital Elevation Model (DEM), and 2024 rainfall data from the Indonesian Meteorological, Climatological, and Geophysical Agency (BMKG). Each variable was classified into five categories based on gridcode values to facilitate spatial analysis. The clustering results revealed two main groups, with the first cluster showing higher landslide potential due to a combination of steep slopes, moderate rainfall, and a high level of urban development. This cluster recorded a Silhouette Coefficient value of 0.75, indicating a high level of landslide vulnerability. In contrast, the other cluster represented more stable terrain, with a Silhouette Coefficient of 0.72. This study is expected to serve as a reference for developing disaster risk-based spatial planning and mitigation strategies.